A smaller, faster variant of Suno's Bark text-to-audio model, capable of generating speech, music, and sound effects from text.